Compared with traditional transcription approaches, the language selectivity dimension method requires no transcription, has low computational overhead, low risk of error propagation, and clear advantages in both robustness and efficiency.